Perceptual video quality evaluation using fuzzy inference system

In this paper, we present a perceptual quality metric for evaluating image and video quality. The metric is based on a fuzzy inference system with Takagi-Sugeno´s inference engine. Three visually important factors including visual masking error, blurring distortion and contrast distortion are used as the inputs to the inference system. Through learning algorithm with the subjective test data, the inference system can predict video quality with high accuracy and monotonicity.